Real case, real data

81 nights. 2,916 guests. This is what measuring my own speakeasy taught me.

Not a demo and not a client. This is El Boiler, my club in San Jose, Costa Rica. Open Thursday, Friday and Saturday, run by me, and logged night by night for a year. Three things came out of that data that I would never have seen otherwise.

Ivo Villalobos

Who is talking

I am Ivo Villalobos. I ran El Boiler Speakeasy, a House music club in San Jose. Reservations, the door, the till, the finances and the marketing all went through me. Every Friday at 3 AM I stared at a spreadsheet trying to work out whether the night had actually made money. Revenight is the system I built to stop guessing, turned into a product.

The operation, raw

81
nights operated
2,916
guests served
36
guests per night, average
3
nights per week

From 25 July 2025 to 27 June 2026. Real operating data, not projections.

Finding 1

My best night was Thursday and I opened five of them.

I programmed Thursdays only five times all year, because I assumed the weekend was where the money was. When I sorted the data, those five Thursdays had the highest average revenue per night of the whole operation: 9% above a Saturday, with fewer people and lower cost.

Thu
+9%
revenue vs Saturday
Fri
33.4
guests per night
Sat
38.6
guests per night

A year of intuition missed it. A table showed it in ten minutes.

What is your Thursday?

Finding 2

The slump lasted seven months and the till never showed it.

August to October averaged 46 guests a night. November to May dropped to 26: down 44%, sustained for seven months. You do not see that in the till. It feels like "this month was slow". On a chart it is a wall.

Guests per night, month by month
34
Jul
43
Aug
48
Sep
47
Oct
26
Nov
11
Dec
36
Jan
31
Feb
27
Mar
24
Apr
26
May
47
Jun

December has a single operated night. Every other month is complete.

What you do not measure per night, you discover half a year in.

Finding 3

Half the people who said yes never showed up.

Across every logged reservation, 51% ended as a no-show. The cause was no mystery: we handed out too many comps to fill the room, and the DJ guest lists came in inflated. We were staffing, stocking and buying for a room that did not arrive.

51%
of logged reservations ended as a no-show

Giving away entry to fill the room is the most expensive way to stay empty.

What I did by hand

  • Weekly newsletter with the event calendar
  • Automated promo 10 days before each birthday, with a gift
  • WhatsApp broadcast lists, sent one by one
  • A signup form that created the contact and pushed it to the list
  • Closing the till and reconciling the night, manually

What the system does now

  • A 0 to 100 score per night, against your real break-even
  • An AI briefing after every night: what worked, what to change
  • WhatsApp reservations answered 24/7
  • Door check-in that builds the guest list on its own
  • P&L per night, importing your POS report

What I measured badly, and why it matters

I did not log expenses with any discipline. I barely recorded walk-ins. Birthdays lived in another tool. And the door check-in was not always marked, because at 2 AM nobody wants to be tapping a screen.

If I missed all of that, as the owner who also built the tool, anyone will miss it. That is why Revenight is designed to capture the data without anyone having to remember to write it down.

Why I closed it

El Boiler closed in June 2026. The building itself capped us: with no emergency exits there was no way to raise capacity or operate calmly, and the business had been scraping by for months. It was my call, made with the numbers in front of me. Being able to make it with the numbers in front of me is exactly what I could not do for the first two years.

Let me show you your numbers, not mine

Send me your POS closing report: PDF, Excel, CSV or a photo. I will hand back the same read you see above, run on your operation, with no commitment and nothing to install.